Google's AI Overviews Add a Tracking Tag to 0% of Their Links vs 99.9% for ChatGPT

A marketing manager opens her analytics to see how much traffic AI sends the company. ChatGPT is right there as a named source, with sessions, landing pages and conversions. Google's AI Overviews, which also cite her pages, are nowhere to be found. Their visits are in the report too, mixed into ordinary Google search traffic with nothing to set them apart.

The difference is a tracking tag: a short code added to the end of a link, such as ?utm_source=chatgpt.com, that tells analytics tools where a visitor came from. Engines that add one make their traffic easy to measure. Engines that do not leave businesses guessing how much AI visibility is worth.

The split is total. 99.9% of the links ChatGPT cites carry its tag, and Gemini tags 97.3%. Google AI Overviews and Perplexity tag none, and Google AI Mode almost none. Because those three engines cite far more sources per answer, they account for 82% of all AI citations.

On September 23, 2026 we asked 2,984 everyday questions to five AI engines from a US location: ChatGPT, Gemini, Perplexity, Google AI Mode and Google AI Overviews. We collected every link cited in the answers, 155,312 in all, including 20,275 links cited by ChatGPT, and checked each one for a tracking tag naming the AI engine.

A tracking tag, or UTM parameter, is text added to the end of a web address that analytics tools such as Google Analytics read to label where a visit came from. A link counts as tagged when it carries a utm_source value naming the engine, such as utm_source=chatgpt.com. An untagged visit is still counted, but under a general label such as Google search or a referral, not under the AI engine's name.

Gemini also tags almost every link, 97.3% of them, with utm_source=gemini. The other three engines leave their links untagged: 0% for Google AI Overviews and Perplexity, and 0.02% for Google AI Mode. For Google's AI features, that means a click from an AI Overview looks the same in analytics as a click from a regular search result.

ChatGPT's tag makes its traffic one of the easiest AI channels to measure. Filter analytics by the source chatgpt.com and you see the sessions that started from a ChatGPT answer, which pages they landed on and whether they converted. The 0.1% of ChatGPT links without a tag show up as direct or unlabelled referral visits.

82% of all AI citations come from Google AI Mode, Google AI Overviews and Perplexity, the engines that add no tag. These engines cite many more sources per answer than ChatGPT, so they account for most citations overall: 127,321 of the 155,312. ChatGPT and Gemini citations, the ones that are easy to identify by their link, make up 18%.

A report built only on tagged traffic therefore sees the engines that cite the fewest sources and misses the ones that cite the most. A business can look as if AI sends it little traffic when most of its AI citations come from Google's AI features, whose visits sit inside organic search numbers.

Set up a ChatGPT traffic report today. In Google Analytics 4 or your analytics tool, create a segment or exploration filtered on session source chatgpt.com, and another for gemini. Together they capture the tagged AI traffic and show which pages AI engines send people to and whether those visitors convert. Compare those landing pages with the pages you want AI to cite; the gap shows which pages need work.

For Google's AI features and Perplexity, use other signals. Track which of your target searches show an AI Overview that cites you, watch Search Console trends for the pages that get cited, follow brand search volume, and read Perplexity visits from the referral report. Suff Digital works as an ai seo company that combines tagged traffic, citation tracking and search data into one AI traffic report.

Embed this research

Cite this study